On the Ecology of Mountainous Forests in a Changing Climate: A ...

166 Chapter 6
However, albeit species compositions differ less the lower the elevation is, there is a
large disagreement on total aboveground biomass between FORCLIM-E/P(/S) and the
other three models at all three sites both under current climate (results not shown) and
under the scenarios of climatic change (Fig. 6.6). At Bern, the difference of total aboveground
biomass leads to low PS coefficients although the simulated species composition
is rather similar among the forest models (e.g. between FORCLIM 1.3 and FORCLIM-E/P
PS = 0.57, between FORCLIM-E/P and FORCLIM-E/P/S PS = 0.82). Thus, we may
conclude that the models are sensitive to the formulation of ecological factors especially
when simulating subalpine forests (cf. Fischlin et al. 1994).
Sensitivity of FORCLIM to the uncertainty inherent in the regionalized scenarios
For the sites Bern and Davos, where the different climate scenarios did not lead to large
differences in the simulated forest community, there is also little sensitivity to the uncer-
Ulmus scabra
Quercus robur
Populus nigra
Fraxinus excelsior
Castanea sativa
Precipitation
400
Acer pseudoplatanus
Acer platanoides
Pinus silvestris
Pinus cembra
Picea excelsa
Larix decidua
Abies alba
Cumulative biomass (t/ha)
300
200
100
Po
P+
P–
0
T– To
T+
Temperature
Fig. 6.7: Effect of the uncertainty inherent in the regionalized climate scenario (Tab. 6.3) at the
site Bever on the steady-state species composition as simulated by the forest model FORCLIM-
E/P.
Symbols: T 0 , P 0 : Best estimate change of temperature and precipitation (Tab. 6.2). T±, P±:
lower and upper end of uncertainty range for temperature and precipitation, respectively (X± =
X 0 ± 2·σ x , where X ∈ {T,P}; cf. Tab. 6.3).